REPORT OF THE SECRETARY. 89 



Two silver-disk pyrheliometers were standardized for the proposed 

 South American expedition. 



Considerable work was done on the apparatus mentioned last year, 

 designed to measure the constant of the fourth power radiation 

 formula. Owing to trouble found in maintaining a vacuum in the 

 apparatus no actual determinations were made. 



Much attention was devoted to the preparation of the equipment 

 of a solar-constant expedition for South America. The purpose of 

 the expedition, as stated last year, is by cooperation with Mount 

 Wilson to secure daily values as far as possible throughout the year 

 for several years, and thus to investigate the influence of solar vari- 

 ation on terrestrial temperature. Many improved devices were in- 

 vented and constructed for the expedition. Among them is a new 

 vacuum bolometer of very high sensitiveness and in every way exem- 

 plary behavior. This instrument is constructed in such a way as to 

 be sealed off when highly exhausted, like an X-ray tube. Having no 

 cocks or windows it requires no further attention to maintain a 

 vacuum indefinitely. The construction of the sensitive strip follows 

 the indications of mathematical analysis covering the whole theory 

 of the bolometer, so that a maximum sensitiveness is obtained. A 

 similar instrument was prepared also for Mount Wilson work. The 

 high sensitiveness of the new bolometer is indicated by the statement 

 that when used with the same spectroscope and galvanometer em- 

 ployed in our Algerian expedition of 1912 more than tenfold de- 

 flections on the solar spectrum were observed with similar conditions. 



Another new instrument is a special machine designed to aid in 

 reducing spectrobolometry, in solar-constant work. Heretofore we 

 have plotted, on large cross-section paper, logarithms of observed 

 radiation against the air masses traversed by the solar beam. Nearly 

 40 such plots, each of six points, are required to represent a morning's 

 spectrobolometry. The plotted points fall in approximately straight 

 lines, whose projection to the zero of air mass yields logarithms of 

 intensities as they would be observed outside our atmosphere. The 

 inclinations of the representative straight lines give the logarithms 

 of the atmospheric transmission coefficients. What I desire to point 

 out is that the process requires taking out about 300 logarithms, 

 besides plotting and extrapolating. 



In the new instrument as shown in the illustration six 16-inch 

 slide rules are arranged to be set at chosen places and at right angles 

 to a horizontal linear scale of air masses. The observations are set 

 up by reading the crossline of the sliders against the central movable 

 slide-rule scales, these latter being set with respect to the fixed scales 

 on the sides so as to apply a small correction for sensitiveness of the 

 bolometric apparatus. A stretched wire is then adjusted to fit the 

 six points as thus plotted. On another slide rule fixed at zero air 



